Changes to the threshold policy would likely mean more Tribes applying directly, and data indicates low chance of substantial increases in overall declared disasters.

• There are 37 Declarations involving tribes (non-COVID, non-EM) under the National Delivery Model (2017-2023).
• 23 are independent Tribal Declarations, where a tribal nation is applicant and recipient.
• 14 are State Declarations, where a tribal nation is recipient, but not applicant.
• These are in Regions 4, 6, 7, 8, 9, 10.
• Regions typically see 0 to 1 Tribal declarations in a year, with R6, R8 and R9 having higher trends (0 to 2 or 4 declarations at peak).
• There are few Tribal PDAs between the 100k-250k range, meaning a few more tribes may apply independently.
• Limited data, but no evidence of turndowns for tribes related to a change in the 100-250k range.

Historic Tables and Estimating Future Years

The following tables show historic, estimated, and estimated + substantive increase in trends of direct Tribal Declarations. Data details and methods are summarized in side notes, with full details in the section below. 11 Section 2. Distribution of Tribal Declarations

Prediction:

If tribal application and declaration activity over the last several years remains moderately similar over the next five years, and declarations are made in consideration of a $100k threshold, we may expect:

• ~30 tribal declarations within the next 5 years.
• Regions responding to 0-1 additional disasters per year, beyond their historic trends.

⊕Random samples of events are drawn from the data, each sample representing a potential future disaster year, then aggregated by Region. If a Region shows 0, it means no events prompted a Tribal DR. If a Region shows NA, it didn’t get drawn in the sample, meaning no events impacted that region in the given year.

_________

Context

Existing Tribal Pilot Guidance requires minimum damage threshold of $250K for tribal declarations. 22 Tribal Declarations Pilot Guidance, January 2017, pg 25 This figure also represents an assumed minimum administrative cost for any declaration. Changing to the $100k threshold is an equitable decision that allows FEMA to service Tribes proportionally to the response historically given to states. 33 See Appendix - Previous Analysis Figure

Most Tribes are quite small compared to county and state equivalents. The current policy creates a filtering effect on our data. Impacts to small tribes and geographies (e.g. a hypothetical $60,000 of flooding damage) may affect 70-90% of a community, but not come close to either the $100k or $250k thresholds. 44 Survivorship Bias Changing the declaration impact threshold doesn’t change the frequency of impacts to tribes, however, the “unknown data” (tribes that may have been impacted, but not assessed, and never working with FEMA) is not something our analysis can directly account for. However, we can say that such impacts, if they represent an increase in events, would be quite small in terms of cost, and are likely occur slightly more often than the typical declared disasters we see for tribes. In a scientific best guess, this might be 1-2 small tribal declarations per Region, per year, instead of the typical 0-1 declarations that Regions currently see. The following analysis attempts to justify this perspective with available FEMA Data.

__________________

Analytic Focus

• Examine distribution of tribal nations across Regions/States and distribution of decs (tribal or otherwise) across Regions/States
• Examine tribal turndowns between $100k and $250k historically
• Examine events that impacted tribes at a level below $100k that they may not have even requested historically.
• Examine events where a tribe came in as a part of a State dec. but may in the future choose to pursue their own instead.

Independent Tribal Declarations by Region and Year

⊕Tribe as Applicant and Recipient, exclusive of State-DRs. Excludes COVID declarations.

• Not all regions have Tribal Declarations in a year.
• Select Regions may respond to 1-2 Tribal Declarations in a year.
• In a 4 year period, the per year probability of a region seeing a declaration would be around 25%.
• If halving the threshold (250k -> 100k) doubles this probability to 50%, it’s a coinflip chance that (select) Regions will respond to a one Tribal Declaration in a year.
• The most frequent value is ‘0’, for example Region 4 had no independent Tribal independent Decs for 4 years, 2018-2021, but there were State-Decs for hurricanes in this region that included tribes.

Gauging the Future

Taking this Region by Year breakdown, we create a sample set of potential disaster years (labelled 2025:2029), using the following method and assumptions. - Sample 25% of all Tribal Declarations - Aggregate to Region - If a Region doesn’t appear in the sample it wasn’t impacted by a disaster.

# #sampling routine
years = 2025:2029  #year set
years_projected = data.frame(Year = 0, "R4"=0,"R6"=0,"R7"=0,"R8"=0,"R9"=0,"R10"=0)

no_tot <- historic_drs2[1:7,1:7]   #Decs by Region/Year - stripped of 'totals'
no_tot[no_tot==0]<- NA

set.seed(6147) #set to reproduce same tables...
for(year in years){
  yt <- no_tot %>%
    pivot_longer(cols=2:7, names_to = "Region", values_to = "DR_Count") %>%
    slice_sample(prop=.25, replace = T) %>% #sample 25% of the distribution...
    group_by(Region) %>%
    summarise(DR_Count = sum(DR_Count, na.rm = T)) %>%
    mutate(Year = year) %>%
    pivot_wider(id_cols = Year, names_from = Region, values_from = DR_Count, values_fill = 0)

  years_projected <- bind_rows(years_projected, yt) #concatenate table
}

years_projected <- years_projected[2:nrow(years_projected),] %>% mutate(Year = as.character(Year))

This gives us the following probabilistic sample table:

⊕NAs for a Region and Year indicates no Events of significant impact.

____________________________________________

Which Tribes may have applied independently?

• The table below shows the Tribes with Disaster Total Amounts above $100k (and below $1M).
• Declaration Type indicates whether the instance was a State-DR or Tribal (independent) DR.
• Assessing specific impacts to Tribes in a State-DR, there were only 4 instances in the last 7 years where a tribe may have applied independently under the 100k, but not the 250k threshold.

⊕Total_Amount is the total Public Assistance spending listed in Grants Manager.

Similar Distributions

One way of thinking about how many Declarations we may additionally see, is by looking at the distribution of declaration data. We would expect the distribution to look approximately symmetrical, or normal (after a log transformation). Where we see gaps is an indication of potential impacts that could have occured (or may occur) but are not currently declared because of the threshold policy.

Here is the distribution plot of current DRs in Grants Manager, data from seven years of Declarations.

And here is a second plot with added bars to fill in the missing gap that should be there for Tribal specific events.

The additions to this plot represent a sample of 4 to 12 small scale impacts, ranging from $22,500 to $130,000, that could occur over a 1-7 year time period. Additional years would also likely add Declarations in the rest of the Tribal distribution (the current range of $200,000-$1M+) Some caveats: - Assumes the $250k threshold is what ‘prevents’ smaller impacts from being registered as Declarations, as opposed to state (or tribal) resources dealing with this impacts without FEMA. - Assumes ‘similar underlying’ character of geographic entities, between Tribes and States, with tribes being a ‘scaled down’ equivalent to States. 55 See Appendix - Fig. 1

To try and validate this idea that there are impacts below the 250k threshold that could have been declared but were not, we now turn to Preliminary Damage Assessment data.

Tribal PDA Amounts

⊕PDA Data 2012-2022, Total Cost Estimate is the sum of all assessed Categories of Work (A-G). The majority of Tribal PDAs have a value of $0 - indicating missing/incomplete or not performed PDAs. These are likely Tribes declared along with a state, or represent PDAs from a COVID period (2020 declarations).

Turndown Status

⊕There is a single example of a Tribes that completed a PDA receiving Turndown Declaration status.

Tribes Near Threshold

⊕Data was filtered to a range between $50,000 -> $250,000.

• There are 4 instances of Tribal PDAs that are above (or near) $100k that were declared.
• These represent specific Tribes that could have applied directly, instead of under a State.

________

Appendix

Current Tribal PDAs

• Shows active Tribal PDAs in the FACT Portal.
• Three instances: 1 Tribal Specific Declaration, 2 Sub-State Declarations
• One state level cost estimate is above the $100k Threshold (NY), tribe name unlisted.

_________________________________________

Previous Threshold Analysis, Key Figures

• State DRs and Tribal DRs are propotional rescalings of one another.
• Tribal incidents being around a tenth of the size of State incidents.

_________

PDA Data and Overall Costs

• Relationship Between PDA and Final Cost is consistent for Tribes and Counties.
• Tribal DRs and Non-Tribal DRs
• Similar linear relationship.